AI, CT, and connected screening ecosystems are redefining the security lane. With automated detection, modular architectures, predictive maintenance, and biometric-linked passenger journeys, airport X-ray scanners are evolving into intelligent platforms that reshape how airports manage risk, efficiency, and compliance.

By: Mirza Bahic; E-mail: mirza.bahic@asmideast.com

For years, the X-ray scanners at the checkpoint symbolized a compromise. It was a compromise between security and convenience, and between safety and speed. Passengers unpacked their belongings into plastic trays, queues moved through terminals, and operators stared at flat, two-dimensional silhouettes for hours on end. The process felt fixed and was shaped by limitations that appeared non-negotiable. That age-old compromise is now being rewritten thanks to a new generation of airport X-ray screening technologies.

Rewriting the Architecture of Airport Screening

Across the globe, airports are deploying a new generation of scanners built around 3D computed tomography (CT), deep-learning-based automatic threat recognition, and tightly connected smart-airport ecosystems. The combined result is not merely gradual improvement but a shift in the very architecture of screening. What once required disassembly, slow belt movement, and focused operator attention can now be achieved with higher accuracy, less friction, and dramatically improved passenger flow.

Regulators, too, are reshaping this part of the security landscape, pushing standards toward higher detection performance while allowing new operational freedoms, such as leaving liquids and electronics inside the bag at airports equipped with certified CT technology. In parallel, regulatory evolution is accelerating the technological one.

To understand where airport screening is heading, we examine the perspectives of three industry manufacturers, Smiths Detection, Gilardoni, and Nuctech, through a unified thematic lens. Their experience, placed against the wider global context, reveals not only how far the sector has advanced, but the direction of the coming decade. At the heart of that evolution is a shift in who, or more precisely what, performs the first layer of screening.

AI Becomes the Primary Screener

First things first, the most significant change in airport screening is not visible to passengers. It lives inside the software stack of modern scanners, where threat detection is increasingly performed by algorithms rather than by the operator’s eye alone.

Juergen Kappler, Portfolio Director of Aviation & Critical Infrastructure at Smiths Detection, describes this approach as two-fold. On one side, deep-learning models trained on very large volumes of X-ray imagery learn the shapes, textures, and visual patterns associated with prohibited items.  On the other hand, more classical image-processing algorithms analyze material properties such as density and effective atomic number in order to identify explosives and similar substances.

These two families of algorithms work together in systems like iCMORE, the company’s automated prohibited-item detection suite for aviation. When deployed with Smith’s CT scanners, iCMORE supports an alarm-only viewing concept. “Only bags flagged as containing potential threats are presented to security operators, while non-alarmed bags move swiftly through checkpoints without manual inspection,” says Kappler. The system is updated continually with new data from regulators and airport stakeholders in order to keep pace with changing threats and test regimes.

In hold baggage, Smiths Detection continues to rely on its long-established explosives detection algorithms, which are still the backbone of ECAC and other certifications. Those algorithms are not static – they are being tuned for current standards such as ECAC EDS 3.2 and are already being developed toward emerging concepts such as EDS-CB C4-limited for cabin baggage.

Gilardoni approaches AI from a slightly different angle. It comes from a background of building X-ray systems where image quality has always been central. Riccardo Bianchi, Product Manager of the Security Business Unit at Gilardoni, notes that the company has been “at the forefront of innovation in X-ray imaging technology for many decades now,” including in-house development of X-ray generators. That investment underpins certified detection performance on conventional and dual-view systems at ECAC EDS-CB C1 and LEDS levels. On top of this, Gilardoni adds pattern recognition of typical threat shapes such as knives, guns, and similar objects, and stresses that these algorithms are continuously updated to keep up with changing forms and concealment methods.

Yet, the company is not tackling AI only in the context of new CT platforms. It is explicitly developing AI solutions compatible with existing conventional scanners, so that airports can upgrade detection capability without wholesale hardware replacement. To support operator training and maintain detection proficiency, Gilardoni integrates Threat Image Projection (TIP) software into its systems, allowing simulated threats to be inserted into live screening sessions for continuous competency assessment without disrupting operations.

Another company, Nuctech, describes a deep-learning-based AI layer running on what it calls its Intelligent Inspection Platform. Rather than using generic models, the company trains on a proprietary baggage image database that includes 3D CT data generated through simulation and through curated real-world samples. This dual approach, which combines simulated threat scenarios with ethically sourced real-world data, allows Nuctech to expand its training datasets rapidly. At the same time, it preserves strict data protection standards.

This allows the algorithms to consider not only shape but also material properties derived from CT, such as density and effective atomic number, when identifying explosives, narcotics, and other contraband. Nuctech stresses that these models are validated through independent tests under ECAC and national civil aviation standards. It also operates a formal “Threat Response Cycle” which takes in new threat information and then pushes updated models into the installed base through secure software updates. This flexible R&D framework enables the company to respond to emerging threats within weeks rather than months, ensuring airport security keeps pace with evolving risks.

Across these three perspectives, AI is clearly no longer an afterthought. It is becoming the first filter in the lane, clearing the majority of baggage automatically and handing only a fraction of images to human operators for further scrutiny. The operator’s role is changing from continuous first-line screening to exception handling. But even the most advanced detection algorithms depend on the quality of the underlying X-ray data, which brings the focus back to CT architecture.

Integration into Smart-Airport Ecosystems

It is evident now that modern scanners are conceived not as standalone boxes, but as nodes in a connected system of baggage handling, operations control, IoT monitoring, and analytics.

Juergen Kappler, Portfolio Director of Aviation & Critical Infrastructure at Smiths Detection, emphasizes that his company’s scanners are designed to integrate “seamlessly into smart airport ecosystems,” supporting open architecture and multiple integration options. Its forthcoming ELECTORA platform is described as an open-standards engine for extracting operational and maintenance insight from scanner data. It is intended to scale across fleets and feed advanced analytics, including predictive maintenance and long-term performance trends. Built on open technical standards and designed for scalability, ELECTORA should provide comprehensive reporting and monitoring capabilities with flexible data export options, enabling airports to contribute scanner telemetry to advanced analytics and real-time operational dashboards.

Riccardo Bianchi, Product Manager of the Security Business Unit at Gilardoni, says that this company presents its scanners as operational-technology components in IoT environments. Machines support real-time diagnostics and health monitoring. This means that data can be exported for analysis, and dynamic dashboards are used to supervise machines, track key performance indicators, and integrate with external monitoring tools.

The company also offers web-based remote control and monitoring dashboards for centralized oversight, and its systems can also be retrofitted with sensors for advanced analytics and predictive maintenance frameworks. The company’s Breva 2.0 automated tray return system adds its own data stream, exposing self-diagnostics that can be used for performance tuning and predictive maintenance across dozens of lanes. Currently operating on approximately 80 lanes deployed worldwide, Breva 2.0 is fully customizable and can integrate with both Gilardoni and third-party scanners, providing advanced self-diagnostics that interface with airport IoT systems to enable comprehensive health and performance monitoring.

Finally, Nuctech describes its systems as “intelligent data nodes” that interface with baggage-handling systems, airport operations centers, and other infrastructure. The company’s systems support standard protocols such as DICOS (Digital Imaging and Communications in Security), facilitating seamless integration with airport Baggage Handling Systems (BHS) and Airport Operations Centers (AOC). Scan results can be used to instruct baggage-routing systems in real time, diverting bags that require further screening, while embedded IoT sensors feed predictive-maintenance platforms that seek to detect wear and anomalies before they result in failure. This shift from reactive repair to proactive prevention allows maintenance teams to schedule interventions during off-peak hours. As a result, it minimizes operational disruption while maximizing system availability.

The net effect is that scanners are no longer just detection devices – they are part of a live data fabric that airports can use to manage flow, maintenance, and risk.

A Three-Track CT Future

So, if AI is the brain, CT is the vision system. The way scanners acquire and reconstruct X-ray data has a direct impact on detection performance, false-alarm rates, and throughput, and here the industry is exploring more than one path.

Smiths Detection continues to build on conventional rotating-gantry CT. Kappler points out that “rotating gantry CT systems have reached a high level of maturity, offering hundreds of views per rotation to achieve exceptional image resolution.” The rotating geometry, combined with anti-scatter grids, produces high-quality volumetric images that feed both explosives-detection algorithms and deep-learning models. At the checkpoint, this design underpins the HI-SCAN 6040 CTiX, which combines full 3D imaging with a belt speed of 0.2 meters per second. The graphical interface deliberately uses the same color scheme as the company’s 2D X-ray systems, easing the learning curve for operators.

At the same time, the company is not limiting itself to conventional CT. It has introduced the SDX 10060 XDi, a system that uses X-ray diffraction instead of tomographic reconstruction. This system has no rotating components. Instead, it analyzes diffraction signatures to identify materials based on their crystalline structure. In early deployments, XDi is positioned as a way to further reduce false alarms in fully automated configurations, especially when combined with CT in a system-of-systems approach. Early trial results show promising outcomes. “False alarms can be reduced by a factor of four to five in fully automated configurations,” Kappler states. In the end, this should lower the need for manual bag checks and improve the overall lane throughput.

Gilardoni is also closely watching this architectural evolution. It recognizes that airports are pushing for higher throughput, lower maintenance, and more cost-effective alternatives to rotating-gantry CT. Bianchi notes that non-rotating CT designs can simplify the mechanical structure and potentially reduce lifecycle costs, but is frank about the fact that static architectures still face challenges in representing objects with the realism and fidelity that rotating CT supports today. According to him, the company is “exploring innovative ways to overcome such technological limitations in order to fully exploit the potential of non-rotating CT machines.”

For Nuctech, static CT is already a strategic choice rather than a future option. The company says it has “pioneered a non-rotating static-gantry CT architecture” that eliminates the heavy rotating assembly, reducing weight and footprint while improving reliability. The simplified mechanical structure significantly improves Mean Time Between Failures (MTBF), translating to higher system availability and lower maintenance costs over the equipment lifecycle.

Nuctech also claims that this architecture provides high-resolution 3D images with full 360-degree coverage and no blind spots, while enabling very high conveyor speeds. With XT2100HS, the company states that this design supports throughputs up to 1,800 bags per hour under standard assumptions for bag length and spacing. Static CT thus becomes not only an imaging choice but a maintenance and throughput strategy.

The industry is therefore not converging on a single architecture. Instead, it is using more powerful software to get the most out of both rotating and static hardware, and, in some cases, adding diffraction-based techniques on top. For airports, the choice will increasingly depend on their mix of performance targets, lifecycle costs, and space constraints.

Throughput: Turning Design into Lane Capacity

Yet, all the imaging and AI sophistication in the world is irrelevant if a system cannot keep up with the passenger flow of a modern terminal. Throughput remains the metric that makes or breaks a deployment.

In hold baggage, Smiths Detection gives the classic benchmark: assume an average bag length of eighty centimeters, a gap between bags of twenty centimeters, and a belt speed of half a meter per second. Under those conditions, the theoretical maximum is about 1,800 bags per hour.

Kappler notes that its SDX 10080 SCT platform is designed to allow even smaller gaps between bags than the standard twenty centimeters and that it performs calibration in the background, so there is no need to stop the belt for routine system checks. In practice, this means that the system comes close to the theoretical throughput ceiling in real-world baggage-handling environments. The SCT’s dynamic calibration capability runs seamlessly during operation, eliminating the traditional need for periodic conveyor stoppages and ensuring sustained performance at optimal throughput levels throughout operational hours.

For cabin baggage, the CTiX runs at 0.2 meters per second and, depending on tray length and spacing, can process on the order of 850 to 900 trays per hour. In the typical configuration, CTiX is integrated with an automated tray return system so that when its detection algorithms trigger an alarm, the tray in question is automatically diverted to a recheck position without interrupting the main flow.

On the other hand, Gilardoni relies on its dual-view systems, which often operate in smaller or mixed-technology airports. They are configured for belt speeds up to 0.3 meters per second in checkpoint roles and up to 0.5 meters per second in hold-baggage systems, where the limiting factor is usually baggage handling rather than operator capacity. Bianchi stresses that its workflow logic allows flagged baggage to move to secondary screening without stopping the lane, and that its algorithms are tuned to minimize false alarms, even at high throughput. Operational feedback from Gilardoni deployments highlights reduced false alarm rates and improved system uptime through scheduled and predictive maintenance as key performance improvements valued by security managers.

Nuctech’s XT2100HS, built on its static CT architecture, is also rated up to 1,800 bags per hour in hold-baggage scenarios. The company emphasizes that throughput is as much about process design as it is about belt physics. It describes a three-level, risk-based screening model where AI performs the first pass and automatically clears most bags, a smaller proportion of images is routed to local or remote operators for detailed review, and only a very small fraction is sent to manual inspection. By structuring operations this way, Nuctech argues that the lane can sustain its mechanical throughput even under peak passenger loads. After 6 to 12 months of operational deployment, airport partners usually report significant reductions in false alarm rates and substantial increases in throughput during peak periods. They also highlight a notably improved operator experience, supported by clearer and manipulable 3D images that enable faster and more confident decision-making.

Throughput is no longer determined solely by belt speed or gantry type. It is the outcome of how imaging, algorithms, automation, and staffing are orchestrated together. While airports now focus heavily on throughput and automation, another foundational aspect of screening systems has reached a level of maturity and regulatory stability.

Radiation Safety Reaches Steady State

Although CT and AI are advancing rapidly, radiation safety is a domain where the fundamentals are stable and heavily regulated.

Kappler underlines that all aviation X-ray systems must comply with strict limits on external dose rate, and that independent radiation-safety officers measure equipment before it goes into operation. To keep leakage as low as possible, the company uses high-end solid-state detectors that require less internal X-ray dose to achieve image quality, along with multi-layer curtain systems and shielding. The inevitable leakage that remains “has been reduced to an absolute minimum,” Kappler notes.

Gilardoni leans on its experience in medical imaging, where dose management has long been critical. Bianchi highlights that the company relies on quality components and X-ray generation and detection to keep X-ray intensity at the minimum level compatible with image quality and certified detection performance. Crucially, Gilardoni’s scanners activate the X-ray beam only while luggage is physically present in the detection area. This beam-gating approach reduces overall radiation emission and optimizes energy consumption throughout operational hours. Bianchi also notes ongoing research into shielding materials and architectures and states that residual leakage is already “orders of magnitude below regulatory limits.”

Their competitor, Nuctech, refers to the ALARA principle, i.e., “As Low As Reasonably Achievable,” as the foundation of its approach. It describes how high-sensitivity detectors allow good images to be formed at lower doses, and how multi-layer physical shielding combined with safety interlocks keeps radiation levels outside the device close to the natural background. All of its systems, it says, are tested by national radiation-protection authorities as part of the certification process.

For airport operators, the conclusion is that radiation is not the differentiator it once might have been. All credible systems operate within a narrow band of regulatory limits and are designed to do so over long lifecycles. The competition now is in how effectively that allowed dose is converted into image quality and detection performance. Yet, screening performance is no longer about standalone machines. Airports increasingly evaluate whether a system can evolve over time and not simply comply today.

Platforms, Not Appliances

With CT, AI, and regulatory standards all evolving, airports are increasingly wary of investing in equipment that cannot be upgraded. In line with this, all three manufacturers therefore stress modularity and long-term upgrade paths.

Smiths Detection’s Kappler gives concrete examples. The company’s first-generation hold-baggage CT systems were installed in 2013 and certified to ECAC EDS 3.0. As standards have progressed to 3.1 and now 3.2, those systems have been upgraded via software. The company notes that EDS 3.2 compliance for the SDX 10080 SCT platform can be achieved through software-only changes, thanks to a CT design with sufficient energy resolution and number of views from the outset.

For cabin baggage, the HI-SCAN 6040 CTiX is built so that new EDS-CB algorithms and iCMORE detection modules can be loaded directly onto existing hardware. In effect, the scanner is a platform whose detection capabilities can be extended and sharpened over time. Kappler also points out that the SCT detector array is designed to allow the addition of more detectors in the future, opening the door to higher belt speeds, and that the platform is engineered to host X-ray diffraction modules if future ECAC or TSA standards require CT-XRD combinations. The SCT system also offers the flexibility to configure with or without an integrated Line-Scanner, which is a complementary imaging technology that enhances specific detection capabilities. This should allow airports to optimize for their operational requirements and budget constraints while maintaining a clear path to future capability upgrades.

On the other hand, Gilardoni focuses strongly on compatibility with existing equipment in the field. Its AI solutions are being developed to run on installed conventional X-ray scanners, preserving the airport’s investment in the mechanical platform. The company also designs integration architectures that can combine data from Gilardoni and third-party equipment, recognizing that many airports operate mixed fleets and will continue to do so while transitioning between generations of technology.

Yet, Nuctech’s approach to the modularity aspect has two strands. On the software side, it positions its systems as “software-defined,” meaning that new AI-based detection functions for emerging threats can be deployed via software updates rather than by swapping out machines. On the hardware side, it highlights field-replaceable modules for critical subsystems such as detector arrays and image processors, allowing targeted hardware upgrades where necessary while keeping the overall platform. The stated aim is to keep pace with evolving standards and capabilities without forcing airports into “rip and replace” cycles.

Modularity has thus moved from marketing slogan to procurement requirement. The question is no longer only whether a scanner can meet today’s standards, but whether its design and vendor roadmap make it likely to meet tomorrow’s.

Cybersecurity in Connected Screening Systems

As scanners become networked devices exposing images and system logs to external systems, they also become part of the airport’s cyber-attack surface. All three manufacturers address this head-on.

Kappler describes a multi-layered security architecture in which data is transferred using encrypted protocols, stored behind strict access rights and authorization controls, and protected by regular vulnerability scanning and security testing. Smiths Detection applies operating-system hardening based on industry-standard CIS benchmarks, leverages Security-Enhanced Linux (SELinux) for additional kernel-level protection, and deploys Host Intrusion Detection Systems (HIDS) to continuously monitor for suspicious activity.

At the device level, kiosk modes, disabled auto-start from external media, and strong password policies reduce the risk of local compromise. Automated vulnerability scans and patch management processes ensure that potential security risks are quickly identified and remediated. Customizable security policies then enforce rigorous access controls across the operational environment.

Other approaches are equally valid. According to Bianchi, Gilardoni’s systems follow data-protection-by-design principles. Scanners can send images to external storage over secure channels with cryptographic protection, and access to those images and related metadata is controlled by role-based access, segregation of duties, and password policies. For systems utilizing removable media, Gilardoni implements OEM-managed secure data-erasure processes to ensure that no sensitive information remains post-disposal, addressing the full data lifecycle from creation to destruction.

Next, Nuctech explicitly references alignment with international data-protection regulations, including GDPR. It uses AES-256 encryption at rest and in transit, enforces strict role-based access control with extensive audit logging, and supports configurations with network segmentation and image anonymization. The company emphasizes that clients retain full control over their data, positioning itself strictly as a technology provider and ensuring that airports maintain complete data sovereignty over all screening information collected by the systems. Nuctech also highlights remote-screening architectures in which operators review images in physically separate facilities, reducing the amount of personal data handled at the checkpoint itself.

This focus reflects a broader shift in the industry: security equipment is now treated as critical IT infrastructure as much as electromechanical machinery, and is assessed accordingly by airport IT and cyber teams.

The Hybrid Future of Screening

Looking forward, the three manufacturers outline trend trajectories that, while distinct, strongly overlap in their direction.

Kappler sees his company’s CTiX as the checkpoint hardware platform on which increasingly capable software will move toward fully automated first-level screening. By combining ECAC EDS-CB algorithms for explosives with APIDS algorithms for prohibited items, and by pushing both toward higher standards such as C4-limited and APIDS 2 or 3, the company expects to reduce the share of images requiring human review to a very small minority.

It has already demonstrated proof-of-concept checkpoints where trays are linked to passengers via biometric capture, such as facial recognition or biometric boarding passes. This foreshadows a future in which passenger identity and baggage are associated throughout the screening process for enhanced security and seamless tracking. For future regulatory requirements, such as potential ECAC EDS 4 or TSA 9.0 standards that may demand CT-XRD fusion, Smiths Detection’s platform strategy enables integration of X-ray diffraction modules into existing CT scanners rather than requiring complete system replacement.

For Gilardoni, Bianchi articulates a strategy defined by modular innovation. Current scanners are deliberately designed so that they can host future automated-detection modules and biometric components, rather than requiring replacement. At the same time, cybersecurity remains a cornerstone of this evolution, with ongoing investments in secure software development, cryptographic protocols, and compliance with evolving global cyber standards as systems become increasingly connected and data-rich.

Finally, Nuctech describes a long-term goal of creating a seamless, intelligent, and integrated security process. The company is actively exploring multi-technology fusion approaches, such as combining CT imaging with trace-detection techniques to improve chemical specificity and precisely identify suspicious substances that CT alone may not characterize properly. It is also exploring biometric-baggage linkage that associates passengers with their luggage throughout the security journey.

Pilot programs are testing facial recognition and other biometric identifiers to create a continuous chain of custody. In this model, each bag is digitally linked to its owner from check-in through final loading, which enables enhanced security tracking and differentiated screening based on passenger risk profiles. In this vision, future checkpoints could become fully automated lanes in which bags are screened and routed with minimal human involvement, and security officers shift into remote supervisory roles, intervening mainly on complex or exceptional cases. This transformation would leverage automation and increasingly powerful AI to achieve operator-free screening lanes, redefining security officers as high-level remote supervisors and managers rather than frontline operators.

All three vendors therefore converge on a picture of the checkpoint as a data-driven, largely automated decision engine that is deeply integrated with passenger identity and airport operations, rather than as an isolated X-ray tunnel with a human operator at the center.

The Age of Smart, Upgradable Screening Is Upon Us

The combined perspectives of Smiths Detection, Gilardoni, and Nuctech show an industry that is changing at multiple levels at once. CT imaging is being refined and diversified through both rotating and static architectures. AI-driven detection is moving into the core of the screening workflow, and throughput is increasingly becoming the result of coordinated system design rather than raw belt speed.

For security professionals, airport X-ray scanners can no longer be evaluated as standalone machines bought for a fixed period and then replaced. They are now platforms, embedded in a wider system of systems, whose value is determined as much by their upgrade path, their integration interfaces, and the vendor’s regulatory posture as by the performance of the hardware at the moment of purchase.

The old compromise between security and convenience is not solved overnight, and legacy fleets will remain in operation for years. But the direction is clear: baggage screening is becoming smarter, more connected, and more automated. The strategic question for airports is no longer whether to adopt this new generation of technology, but how quickly, in what configuration, and with which partners they will move into that future.

Compliance as a Moving Target

No matter how sophisticated the technology is, it only reaches the airport floor if it passes regulatory certification. All three companies pay close attention to this.

Juergen Kappler, Portfolio Director of Aviation & Critical Infrastructure at Smiths Detection, notes that his company’s products are certified by ECAC, TSA, ACSTL, and CAAC, among others. He stresses that the HI-SCAN 6040 CTiX was the first scanner certified for an algorithm that allows screening of liquids up to two liters in cabin baggage, and that its iCMORE automated prohibited-item detection solution was the first to receive APIDS Standard 1 approval at the national level.

Riccardo Bianchi, Product Manager of the Security Business Unit at Gilardoni, mentions a range of certifications across Gilardoni’s conventional portfolio, including ECAC EDS-CB C1, ECAC LEDS, TSA, ENAC, and STAC. He frames certification as an ongoing effort rather than a one-off step, with continuous work to align products to current and emerging standards.

Another competitor, Nuctech, states that its systems have achieved approvals from ECAC, STAC in France, the UK Department for Transport, and CAAC in China, and that it participates in the development of international and national standards. In fact, the company sees itself as “a contributor to the definition of tomorrow’s test regimes rather than merely a respondent.” Beyond pursuing certifications, Nuctech has led or contributed to the development of numerous international and national standards, positioning itself not merely as a compliant vendor but as an active participant in shaping the future regulatory landscape of aviation security.

Given the evolving nature of standards, particularly in cabin-baggage CT levels and potential future CT-XRD combinations, this certification posture matters as much as the current approval list. As a result, Airports increasingly ask not only “is this certified now?” but “is this vendor clearly committed and technically able to keep the system certified as requirements change?”